Molecular ablation of transforming growth factor β signaling pathways by tyrosine kinase inhibition: The coming of a promising new era in the treatment of tissue fibrosis

Fibrotic diseases, a therapeutic challenge

There are numerous human diseases, including systemic sclerosis (SSc), pulmonary, liver, and kidney fibrosis, and the newly recognized nephrogenic systemic fibrosis (NSF), that are characterized by abnormal and exaggerated deposition of collagen in the affected organs (for review, see refs.1–6). Tissue fibrosis causes disruption of the normal architecture of organs and ultimately leads to their dysfunction and failure. The extent and rate of progression of the fibrotic process largely determine the course, response to therapy, and prognosis of these diseases. Although their etiology and pathogenesis are diverse and have not been completely elucidated despite intensive investigations, it is apparent that a common feature is the accumulation of abundant fibrous tissue and the presence of large numbers of fibroblasts displaying an activated phenotype. This phenotype is characterized by a notable elevation in the expression of the genes encoding type I and type III collagens and fibronectin, the initiation of expression of α-smooth muscle actin, and the reduction in expression of genes encoding extracellular matrix (ECM) degradative enzymes.

Regardless of the etiologic event, the resulting alterations in the biosynthetic activity of ECM-producing cells are crucial in the pathogenesis of fibrotic diseases. Indeed, it is the persistent activation of the genes encoding various collagens in these cells which distinguishes controlled repair, such as that occurring during normal wound healing, from the uncontrolled fibrosis that is the hallmark of the fibrotic diseases. However, despite numerous recent advances in understanding of the molecular biology of the events responsible for the regulation of genes encoding collagens and other ECM proteins, there is limited knowledge regarding the intricate mechanisms responsible for the pathologic increase in their expression in fibrotic diseases.

The wide spectrum of organs affected by the fibrotic diseases and the large number of individuals experiencing their devastating effects pose one of the most serious health problems in current medicine and represent an enormous burden on health services and resources, causing severe economic consequences. Despite the high frequency and the diversity of organs affected by the fibrotic diseases, there is currently no effective treatment. The recent elucidation of crucial regulatory pathways involved in the fibrotic response has provided a sound basis for the development of novel and effective means of therapy. In particular, the identification of various important intracellular transduction pathways involved in the transcriptional activation of the genes encoding collagens and other proteins responsible for the fibrotic process has been of substantial importance. The delineation of the critical role of transforming growth factor β (TGFβ) in the development of exaggerated tissue fibrosis and the identification of the specific cellular receptors, kinases, and intracellular mediators that participate in the cellular response to TGFβ in the earliest stages of tissue fibrosis have been the most significant achievements. Unfortunately, the current lack of approved drugs capable of modifying these important mechanisms indicates that the recently acquired knowledge regarding the pathogenesis of the fibrotic process has not been translated into effective therapies in this important area of human health.

Regulation of TGFβ function and signaling

Since the discovery of the potent profibrotic and immunomodulatory activities of TGFβ, this growth factor has been recognized as one of the most important molecules in the pathogenesis of SSc and other fibroproliferative diseases, and a plethora of reports have described studies focused on this subject (for review, see ref.7). Three functionally and structurally similar isoforms of TGFβ exist in humans, and they play important roles in embryonic development, immune responses, and regulation of tissue repair following injury. One of the most important effects of TGFβ is the stimulation of ECM synthesis, as evidenced by a remarkable increase in the production of numerous molecules including type I, III, V, and VI collagens, as well as other relevant proteins such as fibronectin and smooth muscle actin, a molecular marker of activated myofibroblasts. TGFβ also decreases the synthesis of collagen-degrading metalloproteinases and stimulates the production of protease inhibitors such as tissue inhibitor of metalloproteinases 1. Small amounts of TGFβ appear to sensitize fibroblasts to its own effects and maintain them in a persistently activated state involving an autocrine mechanism that causes further production of TGFβ. Bioactive TGFβ in a dimeric form binds to a constitutively active serine/threonine transmembrane kinase known as TGFβ receptor type II (TGFβRII). The signaling events that follow are extremely complex, and the number of intracellular molecules and pathways involved in the process continues to expand (for review, see refs.8–10).

The classic pathway of TGFβ signal transduction into the cell nucleus involves the ligand-bound TGFβRII, which recruits TGFβRI and then transphosphorylates it on 3–5 serine and threonine residues in a short (30–amino acid) glycine- and serine-rich regulatory sequence known as the GS region, as shown in Figure 1. Signaling from phosphorylated TGFβRI to the nucleus then occurs through the Smad family of proteins. Either Smad2 or Smad3, 2 of the 5 receptor-activated Smads, binds to the activated TGFβ receptor complex and becomes phosphorylated by the activated TGFβRI at 2 serine residues near their C-terminal end. Phosphorylation allows these proteins to form a complex with the common-mediator Smad, Smad4, which is a cytoplasmic protein involved in the translocation of the Smad complex across the nuclear membrane into the nucleus. Once in the nucleus, Smad3–Smad4 complexes act as transcription factors, binding with the help of intranuclear proteins that act as transcriptional partners to specific DNA binding sites in the promoter regions of target genes and activating their expression. In contrast, Smad2–Smad4 complexes do not appear to bind directly to DNA promoter sites, but instead, exert their effect through other unidentified transcription factors or coactivator proteins.

Figure 1. Transforming growth factor β (TGFβ) signaling pathways critical for the fibrotic response. Illustrated are 2 canonical Smad pathways originating from 2 representative tetrameric receptors. After TGFβ binding, TGFβ receptor type II (TGFβRII) recruits a type I receptor (TGFβRI) and activates it by phosphorylation (P). Activin receptor–like kinase 5 (ALK-5) then specifically phosphorylates receptor-regulated Smad2 or Smad3, whereas ALK-1 phosphorylates Smad1 (Smad5 and Smad8, also activated by ALK-1, are not illustrated). The receptor Smads then complex with the common-mediator Smad, Smad4, resulting in their transport to the nucleus, where they cooperate with other factors to regulate transcription of critical genes, represented here by the genes for connective tissue growth factor (CTGF) and α2(I) collagen. Also illustrated is a noncanonical pathway resulting in the activation of c-Abl. As pictured, imatinib mesylate blocks the activity of c-Abl, effectively inhibiting the fibrotic response, although the events downstream of c-Abl are currently unknown. Imatinib mesylate also blocks the phosphorylation of Smad1, but here also, the pathway is unknown. PI3K = phosphatidylinositol 3-kinase; PAK2 = p21-activated kinase 2; CBP/p300 = CREB binding protein/p300 transcription cofactor.

Numerous recent studies have provided additional details regarding the complex pathways of TGFβ modulation of the expression of ECM genes, and many participating proteins, such as FK-506 binding protein 12, Smad anchor for receptor activation (SARA), Smurfs, coactivators and corepressors, caveolin 1, and others, have been identified. Furthermore, it has become apparent that important TGFβ effects may be mediated by protein cascades independent of Smad2/Smad3 signaling and that these pathways may become activated in a cell-specific and context-dependent manner. For example, it has been shown that there are 7 different TGFβRI molecular species, known as activin receptor–like kinases or ALKs (ALKs 1–7), encoded in the human genome that exert different and often opposing effects or are only functional in specific cell types. The typical TGFβ signaling cascade involving the receptor-activated Smads 2 and 3 is initiated by the ALK-5 TGFβRI, whereas another TGFβRI, ALK-1, initiates activation of the cascade involving receptor-activated Smads 1, 5, or 8. Although it was initially believed that the ALK-1/Smad1, 5, or 8 cascade was activated only by the binding of a member of the family of bone morphogenetic proteins or that it was only active in endothelial cells, it has now been established that TGFβ can also signal through this pathway in fibroblasts (11).

To further add to the complexity of the TGFβ activation and signaling cascades, it has been recognized that there are numerous other non–Smad-mediated events involved in TGFβ functions and effects (12). Of particular relevance to the present discussion is the recent finding that one of these non-Smad pathways results in the activation of the nonreceptor protein tyrosine kinase c-Abl (13), apparently through the sequential action of phosphatidylinositol 3-kinase and p21-activated kinase 2 as shown in Figure 1.

TGFβ as a target of antifibrotic therapies

The crucial role that TGFβ plays in the initiation and progression of tissue fibrosis and the recognition of its participation in the pathogenesis of numerous fibrotic diseases have focused substantial attention on this growth factor as a target for the development of antifibrotic therapies. Indeed, several strategies have been developed to block TGFβ effects, including soluble TGFβRII fragments, decorin, tranilast, TGFβ-neutralizing antibodies, threonine kinase inhibitors, RNA expression inhibitors such as antisense expression vectors, small interfering RNA (siRNA) or blocking oligonucleotides, and reduction of TGFβ gene transcription with pirfenidone. However, despite the intensive investigation in vitro as well as in vivo in various animal models of fibrosis and in some studies in human subjects, these approaches have either not been effective or are still in the experimental/clinical trial stage and are therefore not currently available for clinical use.

Tyrosine kinase inhibitors

Protein kinase inhibitors are a relatively new class of therapeutic agents that are capable of potent modulation of several cellular phenotypes in cancer and in chronic inflammatory diseases. Of particular interest in the present context are results of numerous recent studies demonstrating that the nonreceptor protein tyrosine kinase c-Abl participates in some of the TGFβ downstream signaling responsible for the development of a profibrotic phenotype response in a variety of cell types. The initial observations leading to this important discovery appeared in the hematology literature, with 3 reports describing simultaneously a remarkable reduction of bone marrow fibrosis in patients receiving treatment for chronic myelogenous leukemia with the c-Abl inhibitor, imatinib mesylate (14–16). Several subsequent studies confirmed the marked improvement in myelofibrosis following imatinib mesylate administration.

Imatinib mesylate is the prototypical drug of the novel class of nonreceptor tyrosine kinase inhibitors. It is a phenylaminopyrimidine-derived small molecule capable of exerting a potent inhibition of several tyrosine kinases including Bcr-Abl, c-Abl, c-Kit, and platelet-derived growth factor receptor at micromolar concentrations. Although the initial kinetic analysis of the mechanisms involved in the potent inhibition of Bcr-Abl kinase by imatinib mesylate indicated that this drug binds to the ATP-binding pocket of c-Abl and efficiently blocks its tyrosine kinase activity, further studies have shown that the binding occurs near the ATP-binding pocket, resulting in a profound conformational change of the kinase active site domain, stabilizing it in an inactive conformation.

Following initial reports describing the inhibition of bone marrow fibrosis by imatinib mesylate, several investigators examined the effects of the drug on in vitro and in vivo models of fibrotic diseases. A landmark study by Daniels et al (17) examined the effects of imatinib mesylate on TGFβ signaling in fibroblasts and on the development and progression of bleomycin-mediated lung fibrosis in mice. The results showed that TGFβ stimulated c-Abl kinase and that imatinib mesylate abolished the fibrogenic response of fibroblasts to exogenous TGFβ in a Smad2/Smad3-independent manner. Furthermore, they showed that in vivo administration of imatinib mesylate in mice abrogated the biochemical and morphologic changes of bleomycin-induced lung fibrosis.

Several subsequent studies explored in further detail the mechanisms of TGFβ signaling inhibition by imatinib mesylate and confirmed the potent effects of the drug in abrogating experimentally induced fibrotic processes such as liver, renal, and myocardial fibrosis (18–20). A more recent in vitro and in vivo study examined the effects of imatinib mesylate on collagen gene expression by dermal fibroblasts cultured from patients with SSc and demonstrated that c-Abl appeared to have a novel and important function in TGFβ-induced fibrotic responses. Furthermore, that study showed that c-Abl inhibition with imatinib mesylate abrogated the development of cutaneous fibrosis induced by bleomycin in mice, a process that mimics the skin fibrosis of SSc (21). Similar observations were recently reported by Distler et al (22), confirming the potent inhibitory effect of imatinib mesylate on the development of scleroderma-like cutaneous alterations in mice receiving dermal injections of bleomycin. Although numerous studies have conclusively demonstrated the participation of c-Abl in the regulation of genes associated with the fibrotic process, the molecular events occurring downstream of c-Abl and the intricate mechanisms of its participation in the regulation of gene expression remain elusive. Furthermore, some of the observed effects may be caused by a reduction in the numbers of tissue fibroblasts, given the potent effects of imatinib mesylate on the proliferation of normal and SSc fibroblasts demonstrated recently (23).

The light at the end of the tunnel for patients with SSc and other fibrotic diseases?

In this issue of Arthritis & Rheumatism, 4 interesting articles corroborate the potential importance of imatinib mesylate in the treatment of a spectrum of fibrotic diseases. In the article by Pannu et al (24), studies with skin biopsy specimens from 8 normal individuals and 8 SSc patients and with cultured fibroblasts from these specimens demonstrated activation of a novel pathway of TGFβ signaling in fibroblasts involving ALK-1 and Smad1, leading to the transcriptional activation of the connective tissue growth factor (CTGF; CCN2) gene. These results are in contrast to those described by other investigators who postulated that the excessive tissue fibrosis in SSc was mediated by activation of the ALK-5/Smad3 pathway. Of substantial relevance to this discussion was the observation that this pathway of TGFβ signaling and the consequent activation of CTGF/CCN2 were abolished by the addition of imatinib mesylate to the cultured fibroblasts. The important role of c-Abl in the activation of the ALK-1/Smad1 pathway and in the phosphorylation of Smad1 was further confirmed with a specific c-Abl siRNA, which completely abolished Smad1 phosphorylation and, remarkably, normalized the excessive collagen production in SSc fibroblasts.

Although these findings indicate that Smad1 phosphorylation is a downstream effect mediated by c-Abl, the mechanisms involved are not known. Since phosphorylation and activation of all receptor-activated Smads is accomplished through specific kinase domains of ALK receptors, the participation of c-Abl in the phosphorylation of Smad1 is puzzling and requires further investigation.

The second article, a case report by van Daele et al (25), describes studies performed with lung fibroblasts obtained from bronchial biopsy specimens from a patient with SSc and pulmonary fibrosis in whom lung involvement progressed despite treatment with twice monthly intravenous cyclophosphamide and low-dose corticosteroids. Imatinib mesylate prevented platelet-derived growth factor–induced proliferation of cultured fibroblasts as well as TGFβ stimulation of type I collagen gene transcription in vitro. Significantly, treatment of this patient with imatinib mesylate resulted in improvement in skin tightness, with a reduction in the modified Rodnan skin thickness score (26) from 18 before treatment to 12 three months after initiation of imatinib mesylate, along with stabilization of pulmonary function as well as findings on high-resolution computed tomography (HRCT) scanning.

The third article is a case report by Distler et al (27) showing that 20 weeks of treatment with imatinib mesylate caused remarkable improvement in pulmonary function and HRCT findings in a patient with mixed connective tissue disease associated with rapidly progressive pulmonary fibrosis. The fourth article reports a study by Kay and High (28), who found that treatment with imatinib mesylate decreased fibrosis and resulted in relatively rapid improvement of skin changes and knee joint contractures in 2 patients with NSF. NSF is a recently recognized fibrotic disease characterized by severe cutaneous and visceral fibrosis that occurs almost exclusively in patients with renal insufficiency following administration of gadolinium-containing contrast agents for magnetic resonance imaging. Severe and progressive joint contractures causing disabilities occur frequently in these patients. The patients described by Kay and High showed substantial clinical improvement as well as histopathologic evidence of reduction of collagen biosynthesis following treatment with imatinib mesylate. These 4 articles provide valuable information regarding the novel role of c-Abl in tissue fibrosis and, more importantly, bring closer the promise of “light at the end of the tunnel” for patients affected by these diseases, as stated in a recent editorial by Wollheim (29).

Speculations and caveats

Several promising therapeutic agents have previously been tested in a variety of fibrotic states, but in most instances, with disappointing results. There is great interest and expectation about the possibility that imatinib mesylate, a drug approved for use in chronic myelogenous leukemias and gastrointestinal stromal tumors (GISTs), may also have a therapeutic effect on fibrotic diseases. Currently, there is only limited experience with the use of imatinib mesylate in the treatment of fibrotic diseases. Thus, given the disappointing results of clinical trials of several promising antifibrotic agents previously studied, it is essential that imatinib mesylate and similar drugs be shown to be effective in a large cohort of patients in well-controlled studies prior to their widespread use in the treatment of fibrotic diseases. Imatinib mesylate has been used extensively in the treatment of chronic myelogenous leukemia and GISTs with spectacular results and relatively little toxicity considering the outstanding benefits achieved. However, one or more adverse reactions occurred in a substantial proportion of the treated patients, including congestive heart failure, edema, muscle cramps, diarrhea, anemia, neutropenia, and thrombocytopenia. It must therefore be demonstrated that long-term use of imatinib mesylate is free of substantial side effects in patients with severe fibrotic diseases, especially with regard to its cardiotoxicity, given the reduced cardiac functional reserve of patients with SSc, pulmonary fibrosis, and NSF.

We have reason to be encouraged, but there is only limited knowledge concerning the signaling pathways in which the pertinent tyrosine kinases participate in relation to the stimulation of fibrosis. While it currently appears that many fibrotic reactions, irrespective of initiating events, terminate in a final common pathway susceptible to this therapeutic intervention, this optimistic view must be substantiated.

Acknowledgements

The authors thank Susan V. Castro, PhD, for her assistance in the preparation of the illustration and the manuscript.